Pony AI maintains its 2026 target of 3,000 robotaxis, with per-vehicle economics already at break-even in Guangzhou and Shenzhen

Overview

This report is Goldman Sachs' conference note on Pony AI following Asia Communacopia + Technology, focusing on robotaxi fleet expansion, the business model, and unit economics. Management maintained its 2026 target of 3,000 robotaxis and operations in 20 cities globally, and said Guangzhou and Shenzhen have already achieved per-vehicle economic break-even.

Core views

Goldman Sachs' core view is that scale expansion and cost declines at Pony AI may together push the robotaxi business closer to, or into, a sustainable operating state. Fleet expansion will be driven by both owned and licensed fleets, and the asset-light model should help the company enter more cities more quickly; meanwhile, there is still room for further declines in cost items such as vehicle design, ADK procurement, remote operations, insurance, networking, charging, and parking.

Analysis framework

The report mainly analyzes management discussions, the company's fleet plan, progress in city deployment, the path to improving unit economics, and Goldman Sachs' long-term EBITDA and peer valuation framework. The valuation section uses 2031E EBITDA and a 20x EV/EBITDA multiple, then discounts back to 2026E using an 11.5% cost of equity.

Methodology notes

• Valuation methodsEV/EBITDA valuation

  Using 2031E EBITDA and a 20x EV/EBITDA multiple to derive a 12-month target price, then discounting back to 2026E.

  The report says the ADR multiple is based on the correlation between Pony AI's global peers' EBITDA growth and EV/EBITDA, while the H-share multiple is based on the correlation between EV/EBITDA and the sum of EBITDA growth and EBITDA margin.

• Valuation methodsCOE discounting

  Use an 11.5% cost of equity to discount the forward valuation back to 2026E.

  The cost of equity assumption includes a 6.5% equity risk premium, a 3.0% risk-free rate, and a beta of 1.3.

• factor_profileGS Factor Profile

  Compare stock characteristics across the four dimensions of Growth, Financial Returns, Multiple, and Integrated.

  Growth is based on forward sales, EBITDA, and EPS growth; Financial Returns is based on ROE, ROCE, and CROCI; Multiple is based on metrics such as P/E, P/B, EV/EBITDA, and EV/FCF; Integrated is the composite percentile after adjusting for growth, returns, and valuation.

• corporate_eventM&A Rank

  Goldman Sachs uses an M&A rank from 1 to 3 to assess the probability of a company becoming an acquisition target.

  Rank 1 represents a 30%-50% high probability, rank 2 represents a 15%-30% moderate probability, and rank 3 represents a 0%-15% low probability; rank 1 or 2 may be incorporated into target price factors.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• PONY AI INC (PONY.US)

  The covered stock in the report; Goldman Sachs rates it Buy and sets a 12-month target price.

  Strengths

  The 2026 target of 3,000 robotaxis is clear, and deployments are already in place across multiple cities in China and overseas; per-vehicle economics in Guangzhou and Shenzhen have already broken even; the asset-light licensed fleet model supports rapid expansion.

  Weaknesses

  The business is still in the early stage of scaling, and profitability depends on continued cost reduction and higher fleet utilization; the current target price mainly relies on forward 2031E EBITDA assumptions.

  Comparison

  The valuation multiple is benchmarked against the correlation between EBITDA growth and EV/EBITDA among Pony AI's global peers.

  Risks

  Slower-than-expected fleet expansion in China and overseas, regulatory changes, slower-than-expected business-model migration, and intensifying competition.

Key data

• Report date2026-05-20Goldman Sachs Equity Research was published at 11:12 AM HKT on 20 May 2026.
• Target fleet size3,000 robotaxisManagement reiterated that this scale will be reached in 2026.
• Target number of cities20 cities globallyManagement expects the global operating footprint to continue expanding.
• Deployed cities in ChinaBeijing, Shanghai, Guangzhou, Shenzhen, Changsha, HangzhouThe current robotaxi deployment cities listed in the report.
• Deployed cities overseasSeoul、Singapore、Doha、Dubai、Zagreb、LuxembourgThe overseas deployment cities listed in the report.
• Cities already at break-evenGuangzhou, ShenzhenManagement said per-vehicle economics have already reached break-even, and margins continue to improve.
• 12-month target priceUS$30.0 / HK$234Based on 20x EV/EBITDA and 11.5% COE discounting.
• Disclosed priceUS$8.32 / HK$64.15ADR and H-share prices shown in company-specific regulatory disclosure.

Impact & implications

If Pony AI can expand its fleet as planned while maintaining declines in unit costs, the commercialization path for robotaxi will become more verifiable. The market may focus more on the speed of city expansion, the rollout of the licensed model, and whether the break-even achieved in Guangzhou and Shenzhen can be replicated in more cities. The valuation upside mainly comes from long-term EBITDA growth and margin improvement at scale, but the high target price also implies significant execution risk.

Risks

• Fleet expansion in China and overseas markets may be slower than expected.
• Regulatory risk could affect robotaxi deployment, operating permits, and the pace of city expansion.
• The shift toward a more efficient or asset-light business model may proceed more slowly than expected.
• Intensifying industry competition may compress prices, utilization, or margins.
• The forward valuation depends on 2031E EBITDA and a 20x EV/EBITDA multiple; if long-term profitability falls short, the target price may be revised downward.

What to watch

• Quarterly progress toward the 2026 target of 3,000 robotaxis.
• Operating permits, rollout speed, and fleet scale for new cities in the global expansion to 20 cities.
• Whether more cities beyond Guangzhou and Shenzhen achieve per-vehicle economic break-even.
• Whether cost items such as ADK, base vehicles, remote driving, ground support, insurance, networking, charging, and parking continue to decline.
• Changes in the mix between owned fleets and licensed fleets, and whether the asset-light model accelerates service expansion.
• The impact of regulatory policy and the competitive landscape on the pace of robotaxi commercialization.